Introduction: Therapy-related Acute Myeloid Leukemia (t-AML) is defined as a type of Acute Myeloid Leukemia (AML) that arises from the mutational consequences of prior cytotoxic, radiation, immunosuppressive, or combination therapies for unrelated diseases. Approximately 5-15% of AML patients have t-AML. Advances in chemotherapies and improved survivorship have expanded the pool of patients at risk for t-AML. There is limited data on temporal trends in survival, length of stay (LOS), and inpatient mortality, which highlights the importance of the current study.

Methods: A retrospective cohort study was conducted using the National Inpatient Sample (NIS) database from 2016 to 2021 to evaluate differences in clinical trends and outcomes comparing t-AML and de novo AML. Cross-sectional data from all AML-related hospitalizations were analyzed. ICD-9 and ICD-10 codes were used for AML. A validated proxy approach - AML with secondary diagnosis codes of history of malignancy, chemotherapy, and radiotherapy - was used to identify t-AML since it lacks accurate ICD codes. Baseline characteristics such as age, sex, LOS, total hospital charges, and In-hospital outcomes such as mortality, sepsis, and acute kidney injury (AKI) were analyzed. Cytopenias and palliative care consults were not included due to a lack of consistent documentation. Multivariable logistic and linear regression models were used to assess the association of t-AML with clinical outcomes.

Results: From 2016 to 2021, we identified a total of 460,090 patients with AML, among which 164,750 (35.8%) had t-AML and 295,340 (64.2%) had de novo AML. 74,785 (45.4%) of t-AML and 130,500 (44.1%) of de novo AML are females. The t-AML hospitalizations were associated with shorter LOS (-1.25 days; 95% CI: -1.49 to -1.02) and lower total charges (-$40,516; 95% CI: -$45,229 to -$35,804) compared to de novo AML. In the adjusted analysis, t-AML was associated with lower in-hospital mortality (adjusted OR 0.54; 95% CI: 0.51-0.57), lower odds of sepsis (adjusted OR 0.59; 95% CI: 0.56-0.61), and lower odds of AKI (adjusted OR 0.59; 95% CI: 0.57-0.62). The proportion of t-AML hospitalizations declined slightly from 0.37% in 2016 to 0.34% in 2021. The in-hospital mortality for t-AML decreased from 6.1% to 4.5%.

Discussion: Our current cross-sectional analysis of AML-related hospitalizations challenges the conventional assumptions about t-AML by demonstrating lower in-hospital mortality, shorter length of stay, reduced odds of sepsis, and AKI compared to de novo AML. Historically, t-AML has been associated with poor prognostic features such as older age, adverse cytogenetics, and chemoresistance. These observed differences may reflect changes in current clinical practice, including the use of lower-intensity therapy, outpatient-friendly regimens, and improved supportive care strategies. The t-AML patients are often closely monitored due to cancer survivorship status, leading to early detection and elective admissions. De novo AML patients are often hospitalized for intensive induction chemotherapy, contributing to inpatient mortality and increased costs.

This study only captures hospitalized patients, and it doesn't reflect overall disease-specific survival or outcomes of non-hospitalized patients. Moreover, the NIS study doesn't include biological and therapeutic heterogeneity. Nonetheless, this data provides a snapshot of real-world care delivery and highlights the evolving clinical profile of t-AML patients. Further studies are needed to assess survival beyond hospitalization and explore disparities in treatment access. These findings suggest a shifting clinical trajectory for t-AML and underscore the need to revisit its classification and management in the real-world setting.

Conclusion: Our large national inpatient analysis suggests that t-AML may no longer uniformly predict worse in-hospital outcomes. These findings likely reflect evolving patient selection, supportive care strategies, and hospitalization patterns. This highlight the need for prospective studies incorporating biological and therapeutic variables to better understand the heterogeneity within t-AML.

This content is only available as a PDF.
2025
Sign in via your Institution